Fix offset when using synthetic noise with BioCAM's sparse compression

[b-grimaud]

Simple fix for #1743

[zm711]

Thanks @b-grimaud. I think this makes sense since we need the offset. We are planning a release in the next couple days, so we will need to discuss this before merge. It may or may not make it in this release, but either way I think it would be more helpful if you put a comment in explaining why we need the offset of 2048 and you can even put in your github issue. something like

# offset of 2048 necessary b/c ....
corrected_data = data + 2048
return corrected_data

[zm711]

We will have @alejoe91 doublecheck this math since he worked with you on the last set of PRs. So you can wait to add the comment until he says its the right idea or just add the comment now before he reviews.

[b-grimaud]

I added some context, hope it's clear enough !

[zm711]

I added some context, hope it's clear enough !

That looks perfect to me. Like I said Alessio will have final say after review, but the logic makes sense to me.